Download UV Spectroscopy of Star-Grazing Comets within the 49 Ceti Debris

Draft version May 13, 2016
Preprint typeset using LATEX style emulateapj v. 05/12/14
UV SPECTROSCOPY OF STAR-GRAZING COMETS WITHIN THE 49 CETI DEBRIS DISK
Brittany E. Miles
Dept. of Physics and Astronomy, University of California, Los Angeles, 430 Portola Plaza, Box 951547, Los Angeles, CA 90095, USA
arXiv:1511.01923v3 [astro-ph.EP] 12 May 2016
Aki Roberge
Exoplanets & Stellar Astrophysics Lab, NASA Goddard Space Flight Center, Code 667, Greenbelt, MD 20771, USA
and
Barry Welsh
Eureka Scientific, 2452 Delmer, Suite 100, Oakland, CA 96002, USA
Draft version May 13, 2016
ABSTRACT
We present analysis of time-variable, Doppler-shifted absorption features in far-UV spectra of the
unusual 49 Ceti debris disk. This nearly edge-on disk is one of the brightest known, and is one of the
very few containing detectable amounts of circumstellar gas as well as dust. In our two visits of Hubble
Space Telescope STIS spectra, variable absorption features are seen on the wings of lines arising from
C II and C IV, but not for any of the other circumstellar absorption lines. Similar variable features
have long been seen in spectra of the well-studied β Pictoris debris disk and attributed to the transits
of star-grazing comets. We calculated the velocity ranges and apparent column densities of the 49 Cet
variable gas, which appears to have been moving at velocities of tens to hundreds of km s−1 relative
to the central star. The velocities in the redshifted variable event seen in the second visit show that
the maximum distances of the infalling gas at the time of transit were about 0.05 to 0.2 AU from the
central star. A preliminary attempt at a composition analysis of the redshifted event suggests that
the C/O ratio in the infalling gas is super-solar, as it is in the bulk of the stable disk gas.
Keywords: protoplanetary disks — comets: general — stars: individual (49 Ceti)
1. INTRODUCTION
A debris disk is a type of circumstellar (CS) disk that is
primarily composed of dust grains coming from planetesimals analogous to comets and asteroids in the Solar System (the Solar System in fact hosts a debris disk, referred
to as the zodiacal dust). In contrast to younger protoplanetary disks, debris disks contain modest amounts of
dust and are optically thin. They are typically discovered
via infrared (IR) photometry of excess thermal emission,
as the CS dust absorbs short-wavelength light from the
central stars and reprocesses it to longer wavelengths.
49 Ceti is a nearby (61 pc) young A1V star that hosts a
bright debris disk containing an unusually large amount
of carbon monoxide gas (Dent et al. 2005; Hughes et al.
2008). The presence of both gas and dust grains in the
disk led to the idea of 49 Cet being a late-stage transitional disk, just on the verge of becoming a typical gasdepleted debris disk (Hughes et al. 2008). Further work
constrained the age of 49 Cet by associating it with the
Argus young moving group (Zuckerman & Song 2012).
This indicates that the age of 49 Cet is ∼ 40 Myr, making it unlikely that the gas in the disk is a remnant left
over from stellar formation, but rather is constantly replenished. Zuckerman & Song (2012) attributed the gas
to frequent collisions of icy comet-like bodies in the disk.
More information for the comet collision scenario has
come from far-IR spectra of 49 Cet obtained with the
Herschel Space Observatory (Roberge et al. 2013); in
contrast to far-IR observations of protoplanetary disks,
C II emission was detected but O I emission was not.
More recently, far-ultraviolet (far-UV) spectra of 49 Cet
showed many strong atomic absorption lines arising from
circumstellar gas (Roberge et al. 2014). Analysis of those
lines indicated a super-solar abundance of carbon to oxygen in the main disk gas, which would be highly unusual
for protoplanetary gas around a nearby main sequence
star like 49 Cet. The far-UV spectra of 49 Cet presented
in Roberge et al. (2014) are also analyzed in this current
paper.
If a large population of comets exists within a disk,
then some of those bodies may get perturbed towards the
host star and sublimate when they get close enough. For
a disk like 49 Cet, which is nearly edge-on from our vantage point (Hughes et al. 2008), some proportion of the
comets may pass through our line of sight to the central
star. During these transits, they would produce Dopplershifted absorption lines superimposed on the stellar spectrum. Such spectral features, which are variable on the
timescale of hours to days, are a long-known and wellstudied aspect of the famous β Pictoris debris disk (e.g.
Beust et al. 1990; Kiefer et al. 2014).
Several other stars have shown Doppler-shifted, variable absorption features in optical spectra of the Ca II
K line, including 49 Cet (Montgomery & Welsh 2012).
However, with the exception of β Pic, no other debris
disk has had multiple UV spectroscopic observations of
the infalling gas done over short time scales. In this paper, we present analysis of infalling gas events seen in
the far-UV spectra of 49 Cet, and strengthen their link
to extrasolar comets by examining the gas dynamics and
composition.
2. DATA
2
1.0
0.8
0.6
0.4
0.2
Visit 1
Visit 2
0.0
CIV
1.5
OI
Flux (10−11 erg s−1 cm−2 Å−1)
Flux (10−12 erg cm−2 s−1 Å−1)
Miles et al.
CIV
1.0
0.5
Visit 1
Visit 2
0.0
1301.0 1301.5 1302.0 1302.5 1303.0 1303.5
Wavelength (Å)
Figure 1. An unvarying O I line showing strong circumstellar
absorption in the 49 Cet STIS data. The Visit 1 spectrum is plotted
with a black solid line, while the Visit 2 spectrum is plotted with
the red solid line.
1548
1550
1552
Wavelength (Å)
1554
1556
Figure 3. The variable C IV lines in the 49 Cet STIS data. The
Visit 1 spectrum is plotted with a black solid line, while the Visit 2
spectrum is plotted with the red solid line. As for the C II lines,
excess redshifted absorption arising from infalling gas is visible in
the Visit 2 spectrum. However, in this case, there is also excess
blueshifted absorption in the Visit 1 spectrum, showing that two
very different star-grazing comet events were recorded.
1.8
2.0
CII
CII*
1.5
1.0
0.5
Visit 1
Visit 2
0.0
1334
1335
1336
Wavelength (Å)
1337
Figure 2. The variable C II lines in the 49 Cet STIS data. The
Visit 1 spectrum is plotted with a black solid line, while the Visit 2
spectrum is plotted with a red solid line. Excess redshifted absorption is clearly visible in the Visit 2 spectrum.
High-resolution (R = λ/∆λ = 228, 000) far-UV spectra of 49 Cet were obtained with the Hubble Space
Telescope (HST ) Space Telescope Imaging Spectrograph
(STIS) on 2013-08-11 and 2013-08-16. A full description
of the dataset appears in Roberge et al. (2014). The radial velocity of the central star is ∼ 12.2 km s−1 (Hughes
et al. 2008). To increase the S/N of the data, each spectrum was rebinned by a factor of 3. The spectrum from
the second visit was interpolated onto the wavelength
scale of the first visit.
Most of the absorption features in the spectra did not
vary between the two visits, showing only stable absorption associated with the main 49 Cet gas disk. A plot
of an unvarying O I absorption line appears in Figure 1.
However, the wings of features arising from C II, C II*
(Figure 2), and C IV (Figure 3) did vary significantly between the two visits. Furthermore, there is one unidentified variable feature in the wavelength range 1594.0 –
1596.5 Å (Figure 4).
In addition to the UV spectra, we obtained optical
spectra of 49 Cet using the Sandiford Echelle Spectrograph on the 2.1-meter telescope at the McDonald Observatory. Five observations were recorded: one on 201308-09, three on 2013-08-11 (the day of STIS Visit 1), and
one on 2013-08-12. The spectra covered the wavelength
Flux (10−11 erg cm−2 s−1 Å−1)
Flux (10−12 erg s−1 cm−2 Å−1)
2.5
1546
1.6
1.4
1.2
1.0
0.8
Visit 1
Visit 2
0.6
0.4
1593.5 1594.0 1594.5 1595.0 1595.5 1596.0 1596.5
Wavelength (Å)
Figure 4. Unidentified variable absorption feature.
region around the Ca II K line (3933 Å) at a spectral
resolution of R ∼ 60, 000. These data are discussed in
Section 4.4.
3. ANALYSIS
For an absorption feature caused by optically thin gas
with complex velocity structure, the apparent optical
depth method can be used to measure (or limit) the
absorbing column densities in particular velocity ranges
(e.g. Roberge et al. 2002). Following Savage & Sembach
(1991), the equation for the column density over a specific velocity range is
Z v2
I0 (v)
me c
ln
dv ,
(1)
N= 2
πe λ0 f v1
Im (v)
where me is the mass of an electron, c is the speed
of light, e is the charge of an electron, λ0 is the central
wavelength of the absorption line, f is the line oscillator
strength, I0 (v) is the intensity of the light from the star
without superimposed absorption (i.e. the continuum),
and Im (v) is the intensity measured after light from the
star has traveled through some medium along the path
to the observer (i.e. the measured spectrum).
Assuming the intensity of the star does not vary on
timescales of days, an increase in the column density
3
Star-grazing Comets in the 49 Ceti Disk
The column density difference ∆N is the material associated with variable gas. This analysis was done for every
variable feature, except the unidentified feature. The 1σ
uncertainties on the ∆N values were determined by propagating the flux measurement errors through Equation 3.
4. VARIABLE SPECIES
4.1. C II
In the Visit 2 spectrum, excess absorption is seen on
one wing of both the C II and C II* lines. Figure 5
highlights this absorption with plots of the Visit 2 spectrum normalized by the Visit 1 spectrum. The excess
absorption is redshifted with respect to the star, showing that the gas was infalling towards the star at the
time of transit. The lower energy levels of the two lines
are not greatly different from each other, and the shapes
of the excess absorption in both lines are also similar.
Therefore, the variable C II and C II* likely arise from
the same parcel of infalling gas. The distinctive triangular shape of the excess absorption is very similar to the
shape of variable atomic absorption seen in UV spectra
of β Pic and attributed to transits of star-grazing planetesimals (e.g. Vidal-Madjar et al. 1994; Roberge et al.
2000).
The large variations in the intensity ratio near v = 0
relative to the star’s velocity are due to the very small
measured intensity values at the bottoms of the saturated
absorption lines from the main disk gas (see Figure 2).
These portions of the data were excluded from our analysis. The C II* line at 1335.7077 Å overlaps with a low
energy absorption line due to Cl I at 1335.7258 Å. However, there is another Cl I line at 1347.2396 Å arising
from the same energy level, which is weak and invariant.
Therefore, C II* is responsible for most of the variable absorption and its column density was calculated using only
the parameters of the C II* 1335.7077 Å line. Table 1
lists all of the variable carbon features analyzed using the
apparent optical depth method, the velocity ranges over
which the variable absorption was significantly detected,
and the total column densities in the variable gas.
4.2. C IV
The two C IV lines show variable absorption on both
the red and blueshifted wings, highlighted in Figure 6.
It appears there was an outgoing event in Visit 1 and
an infalling one in Visit 2. The presence of both red
and blueshifted events is difficult to explain with other
scenarios that do not involve star-grazing comets. For
example, gas accretion should only produce redshifted
absorption, while stellar winds or outflows should only
CII* 1335.7
1.2
Visit 2 / Visit 1
1.2
1.0
0.8
0.6
0.4
1.0
0.8
0.6
0.4
0.2
0.2
−50 0 50 100 150 200
Velocity (km s−1)
0
100 200 300
Velocity (km s−1)
Figure 5. Details of the variable C II absorption events. In the
left panel, the Visit 2 spectrum divided by the Visit 1 spectrum
is plotted, isolating the redshifted excess absorption visible on the
wing of the 1334.5 Å line. The right panel also shows the Visit 2
spectrum divided by the Visit 1 spectrum, isolating the redshifted
excess absorption visible on the wing of the 1335.7 Å line. The xaxis shows the velocity relative to that of the central star. The velocity ranges over which excess absorption is significantly detected
(given in Table 1) are indicated with gray bars.
1.2
1.1
1.2
CIV 1548.2
CIV 1550.8
1.1
1.0
Visit 2 / Visit 1
1.4
CII 1334.5
Visit 2 / Visit 1
v2
I0 (v) I0 (v)
me c
− ln
dv (2)
ln
N2 − N1 = 2
πe λ0 f v1
Im,2 (v)
Im,1 (v)
Z v2 me c
Im,1 (v)
∆N = 2
dv
(3)
ln
πe λ0 f v1
Im,2 (v)
Z
1.4
Visit 1 / Visit 2
along the line of sight will lead to a decrease in observed
intensity at specific wavelengths. The change in column
density can be measured from the ratio of the intensities
measured during and before or after the event, integrated
over the appropriate velocity range.
0.9
0.8
0.7
1.0
0.9
0.8
0.7
0.6
0.6
−600 −400 −200
Velocity (km s−1)
0
0
100 200 300 400 500
Velocity (km s−1)
Figure 6. Details of the variable C IV absorption events. In the
left panel, the Visit 1 spectrum divided by the Visit 2 spectrum is
plotted, isolating the blueshifted excess absorption visible on the
wing of the 1548.2 Å line. The right panel shows the Visit 2 spectrum divided by the Visit 1 spectrum, isolating the redshifted excess absorption visible on the wing of the 1550.8 Å line. The x-axis
shows the velocity relative to that of the central star. The velocity
ranges over which excess absorption is significantly detected (given
in Table 1) are indicated with gray bars.
produce blueshifted absorption. Furthermore, a species
as highly ionized as C IV cannot be produced by photoionization in the CS environment of an A star. As in
the case of the highly ionized, variable species seen in
β Pic spectra, the C IV must be produced by collisional
ionization in hot, dense gas. Such conditions could occur
in the shock at the leading edge of a star-grazing comet
coma (Beust & Tagger 1993).
The red and blueshifted C IV features are badly
blended in the region between the two lines, making
that portion of the data impossible to analyze. Fortunately, both lines arise from the same lower energy level
(0 cm−1 ). Therefore, we could analyze the unblended
blueshifted event in the C IV line at 1548.2 Å and the
unblended redshifted event in the C IV line at 1550.8 Å
to obtain clean measurements of ∆N for both events.
There is a hint of additional blueshifted excess absorption in the Visit 2 spectrum, apparent as a Visit 1 /
Visit 2 ratio greater than 1 near v ∼ −425 km s−1 (see
left panel of Figure 6). This may be a sign of an additional weak outgoing event in Visit 2 simultaneous with
4
Miles et al.
Table 1
Variable features and atomic data
Species
C
C
C
C
II
II*
IV
IV
λ0 a
(Å)
Elower b
(cm−1 )
1334.5323
1335.7077
1548.204
1550.781
0
63.42
0
0
f
c
0.128
0.115
0.190
0.095
Appearance
Visit
Visit
Visit
Visit
2
2
1
2
d
∆v e
(km s−1 )
26 – 143
38 – 211
−372 – −10
13 – 258
∆N f
(1014 atoms cm−2 )
1.145
2.115
1.420
1.603
±
±
±
±
Max. Distance
(AU)
0.380
0.508
0.154
0.196
g
0.17
0.08
···
0.05
a
Rest wavelength
Energy of lower level of transition
Oscillator strength (Morton 2003)
d Visit that shows excess absorption
e Velocity range of excess absorption
f Column density in variable gas
g Maximum distance of gas from star at time of transit
b
c
the strong infalling event. However, as the feature is only
seen in one line and is not robustly detected, no definitive
conclusions can be drawn.
How are the variable gas events seen in C II and C IV
related? In each visit, all spectra were obtained in two
adjacent HST orbits. The first exposure containing the
C II lines was taken about 1.5 hours before the exposure containing the C IV lines. Therefore, the data from
each visit are nearly simultaneous. The similar shape
and velocity range for the redshifted absorption seen in
the Visit 2 data for both species make it likely that the
features are associated with the same infalling gas event.
In contrast, the blueshifted absorption seen only in C IV
during Visit 1 is detected over a wider velocity range, extending to much larger blueshifted velocities. Its higher
velocity suggests the outgoing gas was closer to the central star at the time of transit than the infalling gas event.
This may explain why the blueshifted event was not seen
in the C II lines; the gas may have been hotter and more
highly ionized overall. The different viewing geometry
for the blueshifted event may also have contributed to
its non-detection in C II.
4.3. Mystery Feature
The mystery feature shown in Figure 4 is unlike the
other variable features, since it appears that there is no
stable unvarying absorption at all. This made it difficult to identify the species responsible for the absorption
feature. However, the general shape of the variations
does resemble the variations seen in the C IV lines: redshifted excess absorption in Visit 2 and blueshifted in
Visit 1. This feature is unlikely to arise from Fe II, as
the 1608.46 Å absorption line is seen in the dataset but
does not show significant variation. In far-UV spectra
of β Pic showing variable features, highly ionized species
like C IV generally appear more strongly variable than
less ionized species like C II (e.g. Bouret et al. 2002).
Therefore, the fact that all of the mystery absorption
appears to be variable leads us to suspect that it is due
to a very highly ionized and/or energetic ion.
4.4. Ca II
The optical spectra of the Ca II K line (3933 Å) are
shown in Figure 7. To highlight absorption variations,
the three spectra taken on Aug 11 (the day of STIS
Visit 1) and the spectrum taken on Aug 12 were divided
by the spectrum taken on Aug 9. No significant variation was seen between the first spectrum taken on Aug 11
(Aug 11a) and the Aug 9 reference spectrum. However,
a weak blueshifted absorption feature at v ∼ −8 km s−1
appears in the Aug 11b, Aug 11c, and Aug 12 spectra.
At first blush, this feature could be associated with the
blueshifted absorption feature in the C IV profiles from
Visit 1 (see Figure 3). However, the large difference in velocity shift between the Ca II and C IV variable features
indicates that the low-ionization and high-ionization gas
are not closely associated with each other. It seems likely
that there is a velocity and ionization gradient within
each variable gas event, or that each event may actually be produced by more than one star-grazing comet,
something that has been revealed in ultra-high resolution
Ca II spectra of β Pic (Crawford et al. 1994).
5. DISCUSSION
5.1. Velocity and Stellar Distance
The velocity of the gas as it transits the host star can
be used to find the maximum distance possible of the
comet(s) from the star at a given time. An object that is
gravitationally bound to a star must travel at or below
the free-fall velocity, given by
r
2GM
,
(4)
vf f =
r
where r is the distance of the object, G is the gravitational constant and M is the mass of the central star.
Other massive bodies like planets can speed up or slow
down objects, but for this analysis we considered such
perturbations to be negligible. All other forces on infalling (redshifted) gas, like radiation pressure, will act
to reduce its radial velocity (Beust et al. 1989). Therefore, the free-fall velocity is the maximum velocity that
gas coming from an orbiting body moving towards the
central star can have. Adopting a mass for this A1V star
of 2.7 M⊙ , the fastest velocities for all of the redshifted
events correspond to maximum distances ranging from
0.05 to 0.17 AU, very close to the star.
5.2. Upper Limits on Abundances of Other Atoms
None of the other CS absorption lines arising from the
main disk gas significantly varied between the two visits.
We analyzed unvarying lines arising from several species
(O I, Al I, Si II, S I, Cl I, and Fe II) in order to set
upper limits on the amounts of these gases coming from
star-grazing comets and compare them to the variable
gas abundances. The specific lines analyzed appear in
5
Star-grazing Comets in the 49 Ceti Disk
Table 2. In each case, we placed upper limits using the
strongest line of the species appearing in the dataset.
For this comparison analysis, we analyzed only the
range of velocities over which redshifted (i.e. infalling)
gas was detected in all the variable lines (38 –
143 km s−1 ). This increases the likelihood that we are
analyzing the same parcel of gas for all species. We applied Equation 3 to all the lines, including the variable
ones, and used the flux uncertainties to set 3σ upper
limits on the unvarying gas in this velocity range. The
results appear in Table 2.
To accurately measure total elemental abundances in
the infalling gas, we would need to analyze lines from
many ionization states of each element. This is not possible, since not all of the necessary lines appear in the
data. A measurement of the temperature in the infalling
gas would permit an estimate of the total elemental abundances assuming collisional ionization equilibrium. Assuming local thermodynamic equilibrium, the ratio of
the C II and C II* column densities in Table 2 implies a
temperature of ∼ 1000 K. However, this is a lower limit
on the true temperature in the infalling gas, since there
could be more C II gas in lines at higher energy levels
that we cannot observe. Furthermore, this temperature
is obviously too low to produce C IV by collisional ionization. Nor can we accurately estimate the temperature
in the infalling gas from the abundance ratio of C II and
C IV, since C II may be produced by photoionization as
well as collisional ionization.
However, it appears that C II is a dominant ionization state of carbon in the infalling gas, since no variable C I is seen (although strong CS C I lines appear
in the data) and the variable C IV is about a factor
of three less abundant. In such gas, O I is likely to
be an abundant state of oxygen, since its first ionization potential (13.62 eV) is higher than that of carbon
(11.26 eV). Furthermore, C and O both feel weak radiation pressure from the central A star (Fernandez et al.
2006). These elements should be similarly affected by
dynamical forces and should not be spatially segregated.
Normalized Flux
1.05
Aug 11a / Aug 9
Aug 11b / Aug 9
Aug 11c / Aug 9
Aug 12 / Aug 9
1.0
Table 2
Abundances in the Visit 2 redshifted gas
Species
C II
C II*
C IV
C I**
OI
Al II
Si II
SI
Cl I
Fe II
λ0 a
(Å)
Elower b
(cm−1 )
1334.5323
1335.7077
1550.781
1561.4384
1302.1685
1670.7874
1304.3702
1425.0299
1347.2396
1608.4511
0
63.42
0
43.50
0
0
0
0
0
0
f
c
0.128
0.115
0.095
0.0675
0.049
1.833
0.147
0.192
0.119
0.062
∆N d
(1014 atoms cm−2 )
0.997 ± 0.324
1.654 ± 0.335
0.917 ± 0.008
≤ 0.387
≤ 2.332
≤ 0.017
≤ 2.200
≤ 0.326
≤ 0.480
≤ 0.318
a
Rest wavelength
Energy of lower level of transition
Oscillator strength (Morton 2003)
d Column density in variable gas over velocity range 38 –
143 km s−1
b
c
Therefore, we conservatively chose to represent the total
variable carbon abundance by the sum of the C II ∆N
values (∆NC = (2.65 ± 0.47) × 1014 cm−2 ) and compared
that to the upper limit on the O I abundance.
With these assumptions, the ratio of C to O in the infalling gas is & 1.5, at least 3 times the solar value (solar
C/O = 0.5; Lodders 2003). A carbon overabundance relative to O is also seen in the 49 Cet stable disk gas (C/O
& 4.5; Roberge et al. 2014). Since sub-mm CO emission
is seen from 49 Cet (Hughes et al. 2008), one is tempted
to consider whether CO-rich planetesimals could be responsible for the variable gas. No CO absorption was
seen in the STIS spectra of 49 Cet (Roberge et al. 2014),
likely due to the fact that the disk is not exactly edgeon (Lieman-Sifry & Hughes, in preparation). Obviously,
there is also no detectable CO in the variable gas.
Given the close distances to the star of the variable
gas at the time of transit, any molecular species would
likely be dissociated. If CO was the sole source of the
variable atomic gas, one would expect a C/O ratio close
to 1, which does not appear consistent with the lower
limit on the C/O ratio in the redshifted gas. However,
given the uncertainties about the ionization balance in
the infalling gas, this conclusion is highly tentative. We
note that in the case of β Pic, the measured line-of-sight
abundances of C and CO indicate that CO cannot be the
sole source of carbon in the bulk disk gas (Roberge et al.
2000).
1.0
6. CONCLUSION
1.0
1.0
0.95
−100
−50
0
50
Velocity (km s−1)
100
Figure 7. The Ca II K line in ground-based optical spectra of
49 Cet. The y-axis shows the normalized fluxes, obtained by dividing the spectra taken on 2013-08-11 (Aug 11a, Aug 11b, Aug 11c)
and 2013-08-12 (Aug 12) by the spectrum taken on 2013-08-09
(Aug 9). For clarity, each ratio spectrum was shifted by a constant normalized flux value before plotting. The x-axis shows the
velocity relative to that of the central star. A weak blueshifted
absorption feature appears in the Aug 11b, Aug 11c, and Aug 12
spectra. The variable region is highlighted with the gray bar.
Our observations strengthen the connection between
the variable gas in the 49 Cet disk and star-grazing
comets. First, both red and blueshifted events are seen,
which is hard to explain with other scenarios. Second,
the distance of the gas at the time of transit is within
about 0.2 AU of the central star. The non-detection of
variable O I features suggests that the C/O ratio in the
gas is super-solar, too high for CO to be the primary
source of the gas. Detailed modeling of the ionization
balance in the variable gas will be needed to confirm
this suggestion and to determine the abundance ratios
of other elements. Comets and other small bodies play
an active role in the development of planetary systems,
as we have long known for our own Solar System. As a
young system with an apparently large and active comet
6
Miles et al.
population, 49 Cet is a vital debris disk for learning about
planetesimals in the context of planetary formation and
evolution.
Support for program number GO-12901 was provided
by NASA through a grant from the Space Telescope
Science Institute, which is operated by the Association
of Universities for Research in Astronomy, Inc., under
NASA contract NAS5-26555. A. R. also acknowledges
support by the Goddard Center for Astrobiology, part of
the NASA Astrobiology Institute.
Facilities: HST (STIS)
REFERENCES
Beust, H., Lagrange-Henri, A. M., Vidal-Madjar, A., & Ferlet, R.
1989, A&A, 223, 304
Beust, H., & Tagger, M. 1993, Icarus, 106, 42
Beust, H., Vidal-Madjar, A., Ferlet, R., & Lagrange-Henri, A. M.
1990, A&A, 236, 202
Bouret, J.-C., Deleuil, M., Lanz, T., et al. 2002, A&A, 390, 1049
Crawford, I. A., Spyromilio, J., Barlow, M. J., Diego, F., &
Lagrange, A. M. 1994, MNRAS, 266, L65
Dent, W. R. F., Greaves, J. S., & Coulson, I. M. 2005, MNRAS,
359, 663
Fernandez, R., Brandeker, A., & Wu, Y. 2006, ApJ, 643, 509
Hughes, A. M., Wilner, D. J., Kamp, I., & Hogerheijde, M. R.
2008, ApJ, 681, 626
Kiefer, F., Lecavelier des Etangs, A., Boissier, J., et al. 2014,
Nature, 514, 462
Lodders, K. 2003, ApJ, 591, 1220
Montgomery, S. L., & Welsh, B. Y. 2012, PASP, 124, 1042
Morton, D. C. 2003, ApJS, 149, 205
Roberge, A., Feldman, P. D., Lagrange, A. M., et al. 2000, ApJ,
538, 904
Roberge, A., Feldman, P. D., Lecavelier des Etangs, A., et al.
2002, ApJ, 568, 343
Roberge, A., Welsh, B. Y., Kamp, I., Weinberger, A. J., & Grady,
C. A. 2014, ApJ, 796, L11
Roberge, A., Kamp, I., Montesinos, B., et al. 2013, ApJ, 771, 69
Savage, B. D., & Sembach, K. R. 1991, ApJ, 379, 245
Vidal-Madjar, A., Lagrange-Henri, A.-M., Feldman, P. D., et al.
1994, A&A, 290
Zuckerman, B., & Song, I. 2012, ApJ, 758, 77